Back Propagation Neural Network Applied to Modeling of Switched Reluctance Motor

Author

Sun, Jianbo ; Zhan, Qionghua ; Guo, Youguang ; Zhu, Jianguo

Author_Institution

Dept. of Electr. Machinery, Huazhong Univ. of Sci. & Technol., Hubei

fYear

0

fDate

0-0 0

Firstpage

151

Lastpage

151

Abstract

This paper presents a back propagation neural network (BPNN) application for modeling of SRM, incorporating finite element analysis. Firstly, the magnetic curve of ferromagnetic material is smoothed by a BPNN. Secondly, this paper deduces the formula of magnetic force based on the local Jacobian derivative method and the magnetic vector potential. Thirdly, the determination of the optimal BPNN structures and learning times is introduced. At last, a dynamic model of SRM based on BPNNs is constructed. The validity of the model is proved by comparing the simulation results with the experimental results

Keywords

backpropagation; electric machine analysis computing; ferromagnetic materials; finite element analysis; magnetic forces; neural nets; reluctance motors; back propagation neural network; ferromagnetic material; finite element analysis; local Jacobian derivative method; magnetic curve; magnetic force; magnetic vector potential; switched reluctance motor; Couplings; Jacobian matrices; Magnetic analysis; Magnetic flux; Magnetic forces; Magnetic materials; Neural networks; Reluctance machines; Reluctance motors; Torque;

fLanguage

English

Publisher

ieee

Conference_Titel

Electromagnetic Field Computation, 2006 12th Biennial IEEE Conference on

Conference_Location

Miami, FL

Print_ISBN

1-4244-0320-0

Type

conf

DOI

10.1109/CEFC-06.2006.1632943

Filename

1632943